To investigate the effects of LIPUS irradiation on the rate of OTM and alveolar bone remodeling, OTM models on the bilateral maxillary first molars of rats were constructed; a schematic of the study flow is presented in Figure 1A. LIPUS accelerated OTM from the third day, and there were statistically significant differences in the tooth movement distance on the 15th and 21st days (Figure 1B). Given that the alveolar bone between the mesial and distal roots of the first molar includes both tension and pressure areas during OTM, the changes in the alveolar bone volume and microstructure of the region of interest (ROI) were evaluated using micro-computed tomography (micro-CT). Representative reconstructed images and sectional images in different directions are shown in Figure 1C. The LIPUS group showed more movement compared to the control group, although it was not entirely horizontal. It can be seen that on the 21st day, there was a smaller OTM distance and fewer bone trabeculae in the alveolar cancellous bone, with a clear porous structure on the control side. The quantitative results indicated that LIPUS treatment increased the bone volume/total volume (BV/TV), trabecular number (Tb.N) and trabecular thickness (Tb.th), and decreased the trabecular separation (Tb.Sp) compared with the control group during OTM (Figure 1D). Next, the histological changes in the alveolar bone around the first molar after 21 days of OTM under LIPUS treatment were observed using hematoxylin-eosin (HE) and Masson staining. Similar to the micro-CT results, the LIPUS side had more newly deposited bone at the edge of the trabecula in the alveolar cancellous bone and the bone trabeculae were thick and closely arranged. Higher new bone formation in cortical and cancellous bones of LIPUS side, indicating promising bone formation potential. In addition, the proximal buccal cortex of the orthodontic teeth treated by LIPUS was significantly thickened and suggesting that LIPUS may promote compensatory bone formation during OTM (Figure 1E). Micro-CT analysis showed that LIPUS stimulation slightly increased the cortical bone thickness on the pressure side, although there was no statistical difference in Ct.Th, which may be due to insufficient cortical mineralization during the shorter experimental period (Figure 1D).

To further explore the effect of LIPUS on bone homeostasis during OTM, calcein and alizarin red S (ARS) fluorescent labels were used to analyse the regions of bone formation. As shown in Figure 2A, the fluorescence signal of the LIPUS side was significantly enhanced, and the distance between the two labelled edges was increased. The bone mineral apposition rate (MAR) of the LIPUS side was significantly greater than that of the control (Figure 2B). These results confirm that LIPUS promotes changes in compensatory osteogenesis in the bone cortex on the pressure side during OTM in rats. To test the hypothesis that EphrinB2/EphB4 signaling may be involved in LIPUS mechanical signal sensing and transduction, the expression patterns of EphB4 and EphrinB2 in the periodontal ligament (PDL) and alveolar bone were examined by Immunohistochemistry (IHC). The results showed that EphB4 was expressed in both the PDL and alveolar bone, while EphrinB2 was mainly expressed in the alveolar bone and was rarely found in the PDL (Figure 2C). Quantitative analysis showed that the positive areas of EphB4 and EphrinB2 were more widely distributed in the LIPUS group, although the average optical density of EphrinB2 was decreased compared with the control side, suggesting the effect of LIPUS on alveolar bone remodeling may be achieved by increasing the binding between EphB4-EphrinB2 (Figure 2D). These findings suggest that LIPUS affects the expression of the EphrinB2/EphB4 signal during OTM, and thus, may influence the homeostasis balance of the orthodontic alveolar bone, resulting in compensatory bone cortex formation through osteoblast-osteoclast crosstalk.

BMSC-derived osteoblasts in mono-culture or in co-culture system with BMM-derived osteoclasts were studied using the methods shown in Figure 3A. First, we attempted to determine the effect of LIPUS on the differentiation of osteoblasts. Alkaline Phosphatase (ALP) and ARS staining revealed that, compared with the control, BMSC-derived osteoblasts that were exposed to LIPUS treatment had increased ALP activity and mineralized nodule formation, which is consistent with previous research reports. Interestingly, BMSC-derived osteoblasts in the direct co-culture system with BMMs had significantly increased ALP activity and formation of mineralized nodules following LIPUS treatment, suggesting that BMM-derived osteoclasts may work in synergy with LIPUS to promote the osteogenesis of BMSC-derived osteoblasts. To explore whether BMMs themselves affected the osteogenic potential of BMSC-derived osteoblasts, ALP activity and calcium salt deposition were evaluated in the osteoblast-osteoclast co-culture system without LIPUS treatment. The results revealed no significant changes compared with BMSC-derived osteoblasts in mono-culture (Figures 3B, C). The osteogenic phenotype of BMSC-derived osteoblasts in each group was determined by real time-quantitative PCR (RT-qPCR). The results showed that the mRNA expression levels of the osteogenic specific transcription factor Runx2 and the early osteogenic differentiation marker gene Col1a in osteoblasts were slightly increased by LIPUS treatment, and significantly increased by the combination of osteoclast co-culture and LIPUS stimulation. Meanwhile, there was no significant enhancement in osteogenic gene expression in the osteoblast-osteoclast co-culture system without LIPUS treatment. The mRNA expression of the cell membrane receptor EphB4 was upregulated by LIPUS stimulation, and its expression was significantly increased in the osteoclast co-culture system with LIPUS treatment, while the expression of the cell surface ligand EphrinB2 did not change significantly under the action of LIPUS (Figure 3D). The protein expression levels of EphB4 and EphrinB2 in each group were consistent with the transcriptional levels (Figures 3E, F).

Since EphrinB2/EphB4 signals have bidirectional regulatory effects on the osteoblasts and osteoclasts, we sought to determine the effect of LIPUS on osteoclast differentiation in the co-culture system. LIPUS was applied to the BMMs in mono-culture and in co-culture system with BMSC-derived osteoblasts. Then, tartrate-resistant acid phosphatase (TRAP) staining and F-actin ring fluorescence was used to evaluate the osteoclast differentiation of BMMs. The results showed that LIPUS had no significant effects on the number and size of osteoclasts in the mono-culture BMMs. Regardless of the presence of LIPUS stimulation, the number and size of osteoclasts were increased (Figures 4A, B) and the F-actin ring of the osteoclast skeleton was significantly visible (Figures 4C, D) in the BMM-derived osteoclasts in the co-culture system with osteoblasts. This suggests that the formation and maturation of osteoclasts were upregulated by the combination of BMSC-derived osteoblasts. RT-qPCR results showed that LIPUS had no significant effects on osteoclast differentiation genes in mono-culture BMMs, but slightly increased the c-fos and MMP9 mRNA expression levels of osteoclasts in the osteoblast cell co-culture system. This suggests that LIPUS promotes the differentiation and maturation of osteoclasts in a co-culture system (Figure 4E). At the same time, EphrinB2 mRNA, mainly expressed on the membrane of osteoclasts, was significantly increased in BMM-derived osteoclasts in the osteoblast cell co-culture system compared with the mono-culture BMMs; however, its expression was not affected by LIPUS.

To mimic the specific ligand to EphB4 on the osteoclast surface, recombinant mouse EphrinB2-Fc chimera protein was utilized to study whether EphrinB2 is involved in the regulatory effect of LIPUS on BMSCs osteogenesis. The methods utilised in this part of the study are shown in Figure 5A. ALP and ARS staining were used to evaluate the osteogenic ability of each group, and the results indicated that the osteogenic effect was strongest for the combination of LIPUS and EphrinB2-Fc, followed by the single-stimulus conditions of LIPUS or EphrinB2-Fc (Figures 5B, C). RT-qPCR and Western blotting were used to evaluate the expression changes in osteogenesis-related marker genes and EphB4. Compared with the control group, LIPUS increased the mRNA expression levels of Runx2, Bglap and EphB4 in BMSC-derived osteoblasts, but there was no significant difference in the osteogenesis genes expression by EphrinB2-Fc single treatment. With the combined stimulation of LIPUS and EphrinB2-Fc, osteoblastic differentiation was increased remarkably (Figure 5D). Moreover, the Western blotting results confirmed that EphB4 protein expression was increased with LIPUS stimulation, and EphrinB2-Fc enhanced this upregulation (Figures 5E, F).

To explore the role of EphB4 as an EphrinB2 receptor in LIPUS signal perception and BMSCs osteogenesis regulation, we tested the biological function of EphB4 by knocking it down in BMSCs. ShEphB4-2 was selected due to the higher efficiency of interference through qPCR (Figure 6A). The Western blot and immunofluorescence results confirmed that shEphB4 effectively reduced the protein level of EphB4 (Figures 6B–D). ALP and ARS staining results indicated that knockdown of EphB4 by shRNA led to a decrease in ALP activity and mineralization in BMSCs. However, with the combination of EphrinB2-Fc, LIPUS showed great osteogenic potential, while this effect was remarkably attenuated by shEphB4 (Figures 6E, F). This suggests that the forward signal of EphrinB2-EphB4 may be involved in the effect of LIPUS on BMSCs osteogenesis, but it may not be the only factor contributing to this effect. Evaluation of wound healing and the transwell test revealed that the synergistic effect of LIPUS and EphrinB2-Fc significantly improved the migration ability of BMSCs, while EhpB4 depletion impaired the enhanced migration ability of BMSCs (Figures 6G, H). This suggests that LIPUS may affect the migration of BMSCs through EphB4 signaling, thereby affecting the osteogenesis of BMSCs.

Having revealed the LIPUS promotes the migration and osteogenesis of BMSCs by increasing the expression of EphB4 on the membrane of BMSC-derived osteoblasts, especially under the co stimulation of osteoclasts or EphrinB2-Fc, we further sought to reveal the underlying molecular mechanism. To determine the effect of LIPUS on the interaction between EphrinB2 of osteoclasts and EphB4 of osteoblasts, the co-culture system was treated with LIPUS stimulation and osteogenic medium for 7 days. The immunoprecipitation results revealed that LIPUS significantly increased the binding between endogenous EphrinB2 and the EphB4 protein (Figure 7A). The immunofluorescence of His-Tag-antibody cross-linked with EphrinB2-Fc chimera protein represents the amount of EphrinB2-Fc binding to the EphB4 receptor on the osteoblast membrane; the results indicated that LIPUS facilitated the attachment of EphrinB2-Fc to EphB4 (Figures 7B, D). Previous studies have reported that the cytoskeleton functions as a converter for MSC differentiation by responding to EphB4 then affecting YAP nuclear translocation. To study the structure of the stress fibre response to LIPUS stimulation, F-actin was stained and the fluorescence intensity of YAP in nuclei was quantified. The results indicated that, in the control group, F-actin was arranged along the long axis of the cell in the form of filaments. LIPUS and EphrinB2-Fc promoted the gathering of F-actin filaments around the nucleus and the formation of filopodia and lamellipodia on the cell membrane, while the depletion of EphB4 mitigated the changes in F-actin and YAP caused by LIPUS (Figures 7C, E). These results suggest that EphB4, as a membrane receptor, exhibits enhanced binding with EphrinB2 following LIPUS treatment, and then affects YAP nuclear localization through cytoskeleton rearrangement and downstream target genes. Western blotting for nucleocytoplasmic separation showed that LIPUS slightly increased the overall abundance of YAP in BMSC-derived osteoblasts, significantly promoted the nuclear localization of YAP, and slightly reduced its phosphorylation in the cytoplasm with the coordination of EphrinB2-Fc (Figures 7G, H). Finally, to determine the biological role of YAP in the transduction of the LIPUS- EphB4- EphrinB2- F-actin cascade signal, Verteporfin was added to the cells to block the role of YAP. The results of ALP and ARS staining indicated that YAP blocking reduced the activity of ALP and the level of calcium salt deposition, supporting the speculation that the positive regulatory effect of EphB4 on osteogenesis may be related to YAP signaling (Figure 7F).

Eight-week-old male Standard Deviation (SD) rats (weight 200 ± 20 g) for OTM model and C57BL/6 mice for primary cell isolation and culture were provided by the SJA Laboratory Animal CO., LTD (Hunan, China). All animals were housed under standard conditions: room temperature 22°C ± 2°C, 12 h light-dark cycle, water ad libitum, standard pellet diet. The experimental procedures were approved by the Ethics Committee of the College of Stomatology, Chongqing Medical University (Approval number: CQHS-REC-2022 (LSNo.104)). This study was carried out in compliance with the ARRIVE guidelines 2.0.

Significant intergroup differences of 0.6 mm (standard deviation = 0.1 mm) in control and 0.8 mm (standard deviation = 0.1 mm) in LIPUS were determined on the basis of published research (Arai et al., 2020). With an 80% power, an α of 0.05, and a 2-tailed test, 4 rats per group were recommended. After 1 week of adaptive feeding, 24 SD rats were anesthetised with 2% isoflurane gas inhalation and shallow grooves were carved in the gingival margin of the upper incisor and first molar. Ni-Ti closed coil springs (Shinye, Zhejiang, China) were fixed between the maxillary bilateral first molars and incisors. A force of 50 g was applied to induce mesial movement of the first molar. The orthodontic appliances were checked every day and were reinstalled or replaced in time if the appliances fell out or were damaged.

The LIPUS generator was provided by the National Engineering Research Centre of Ultrasound Medicine, Chongqing Medical University (Chongqing, China). To reduce the impact of individual variation, each rat was exposed to LIPUS on the right side, and the left side was used as self-control with the LIPUS generator switched off (n = 24/group). LIPUS was performed from the first day after OTM. The ultrasonic transducer was placed in contact with the mucosa of the first molar through the couplant and LIPUS was applied at an intensity of 30 mW/cm², frequency of 1.5 MHz, repetitive pulse frequency of 1.0 kHz and burst sine wave of 200 μs. The LIPUS stimulation was administrated for 30 min/d for 21 consecutive days. Three rats in each group were sacrificed with CO₂ at appropriate time points (days 3, 7, 11, 15 and 21), and samples were obtained for histological studies.

For the culture of BMSCs, the femur was separated from male 8- to 10-week-old mice and bone marrow cells were flushed out with α-minimum essential medium (α-MEM) (C12571500BT, Gibco, Grand Island, NY, United States) supplemented with 1% penicillin/streptomycin and 10% fetal bovine serum (FBS) (BBP5, Moregate, New Zealand). The cell medium was changed every two-three days until the cells reached 80% confluence in approximately 7 days. The cells were seeded into plates at a ratio of 1:2 or 1:3. For osteoblast differentiation, BMSCs were cultured in osteogenic medium (α-MEM containing 10% FBS, 1% penicillin/streptomycin, 10 mM β-glycerophosphate, 50 mg/L ascorbic acid and 100 nM dexamethasone) for 7 days.

For BMMs culture, the femur was prepared as described above. The flushed bone marrow cells were cultured in a-MEM for 24 h and the floating cells were collected. The BMMs were obtained by gradient centrifugation with Histopaque (10771, Sigma, St. Louis, MO, United States), and were seeded into plates after resuspension. Osteoclast differentiation was achieved by culturing BMMs (2× 10⁵ per well in a 24-well plate) in osteoclastogenic medium (40 ng/mL M-CSF [CB34, Novoprotein, Suzhou, CN] + 100 ng/mL RANKL [CR06, Novoprotein, Suzhou, CN ]) for 7 days.

The method for building the osteoblast-osteoclast co-culture system has been widely described and analysed (Borciani et al., 2020; Hong et al., 2022). BMSC-derived osteoblasts were seeded at 2 × 10⁴ cells per well in a 24-well plate and BMMs were seeded on the attached osteoblasts (1 × 10⁵ cells per well) in osteoclastogenic medium (40 ng/mL M-CSF +100 ng/mL RANKL+ 10⁻⁸ mol/L 1,25(OH)-dihydroxyvitamin D3) on the following day.

The cells were irradiated with LIPUS (30 mW/cm², 1.5 MHz) in a water bath, consistent with our previous research (Wang et al., 2018; Li et al., 2020; Ying et al., 2020). LIPUS stimulation was performed for 20 min daily for 7–14 days. Cells in the control group were treated with sham LIPUS stimulation (no energy output from the transducers).

In order to stimulate forward signaling, soluble recombinant mouse ephrinB2-Fc chimera (496-EB, R&D, Minneapolis, MN, United States) was preclustered with His-Tag antibody (MAB050, R&D) and used at a final concentration of 1 μg/mL.

Three shRNAs targeting EphB4 (NCBI gene ID: 13846) and non-targeted control shRNA were synthesized (Tsingke, Beijing, CN). The shRNA sequences were as follows:

#1: CCG​G-AGC​CAA​CAG​CCA​CTC​GAA​TAA-CTC​GAG-TTA​TTC​GAG​TGG​CTG​TTG​GCT-TTT​TTT;

#2: CCG​G-CCT​GTG​TCG​AAT​TGG​GTA​TTA-CTC​GAG-TAA​TAC​CCA​ATT​CGA​CAC​AGG-TTT​TTT;

#3: CCG​G-CGG​ATC​TGA​AAT​GGG​TGA​CTT-CTC​GAG-AAG​TCA​CCC​ATT​TCA​GAT​CCG-TTT​TTT.

BMSCs were seeded at a density of 2× 10⁵ in six-well plates and were transfected with pTSB -shEphB4 plasmid using Lipofectamine 3,000 (L3000015, Invitrogen, Carlsbad, CA, United States) within 16 h. For RNA and protein isolation, cells were harvested at 48 h and 72 h after transfection, respectively.

All quantitative data are presented as the mean ± SEM of at least three independent measurements. Differences between two groups were assessed with Student's t-tests and differences between four groups were assessed with one-way ANOVA. All statistical analyses were performed using Prism 8.0 software. Differences at p < 0.05 were considered statistically significant.